4 ms·
No catch at all!! There's 0 approximations, so everything is exact! We just have a custom backprop engine, rewrite everything into OpenAI's Triton language, and
by danielhanchen 3y ago
No catch at all!! There's 0 approximations, so everything is exact! We just have a custom backprop engine, rewrite everything into OpenAI's Triton language, and do all the differentiations and maths ourselves :)
Unsloth can fit 4x longer context windows than HF + Flash Attention 2 as well with our latest long context update, so 30% less VRAM use, at the expense of slightly +1.9% overhead.